Multivariate Analysis Is Appropriate for a Quantitative

¶ … multivariate analysis is appropriate for a quantitative study. Conclude your posting by describing your personal interests in one of the multivariate statistical tests and its potential usefulness to you in future research.

The question of whether to use a multivariate vs. univariate or bivariate statistical tests is rather straightforward. Multivariate statistics provide analyses in cases where there are more than one independent variable and/or more than one dependent variable (Tabachnick & Fidell, 2012). Univariate analyses are statistical tests that involve one variable, whereas bivariate tests/statistics involve two variables such as a Pearson correlation coefficient (Runyon, Coleman, & Pittenger, 2000). It is certainly possible to do many different univariate or even bivariate analyses in the place of a multivariate one; however, the researcher loses statistical power, increases the probability of a TType I error, and is unable to determine all of the relationships in the data that can be analyzed with a multivariate analysis (Tabachnick & Fidell, 2012). Multivariate statistics are extension of univariate and bivariate statistics and represent the general case, whereas univariate and bivariate statistics are special cases of the multivariate model (Tabachnick & Fidell, 2012). Multivariate statistics allow for the simultaneous analyzing of multiple dependent and/or independent variables thus allowing aforementioned advantages in addition to representing real -- world models more accurately where numerous variables often contribute to an outcome or observation (Runyon, Coleman, & Pittenger, 2000).

It is important to note that all of the different types of multivariate statistical techniques have limitations and guidelines as to their use and application. Thus, while multivariate statistical techniques allow for greater freedom in research design and analysis they do not allow the researcher to analyze a limitless number of variables, ignore the restrictions of the data (e.g., continuous vs. discrete data), or make an unlimited number of comparisons. Like any research tool multivariate statistics have their limitations. Different types of multivariate tests allow a researcher to answer different types of questions, but a researcher's ability to make inferences is always limited by the type of data collected in the methodology used to collect it.

I am interested in how gender rates differ in cases of human trafficking into the United States. Currently there is a study that proposes looking at the differences in gender in human trafficking cases based on the data from The National Human Trafficking Resource Center. This particular to the research could be adequately performed using a simple t-test or one-way ANOVA. However, other variables could be added to the methodology and multivariate analyses could offer much more information regarding the differences in males and females involved in human trafficking. Using other variables such as geographical area trafficked to (e.g., in the United States this could be divided into Midwest, North, South, etc.), the type of context into which human victims are placed (e.g., hard labor, domestic work, prostitution, etc.), ethnic background of the trafficked individual, and others allows for a richer analysis. In this case there would be multiple independent variables that can contribute to the differences in the number of males and females that are trafficked into the United States and multivariate techniques such as factorial ANOVA (or its counterpart multiple regression depending on the type of question) is more appropriate. In this type of study we would expect that the particular context would lead to differences in the rates of males and females trafficked into the country. For example, we would expect that people put into hard labor jobs such as